useful design's picture

Alastair (@usefuldesign)

Groups

  • Vuo Founder

Compositions

1 year ago
useful design's picture
Alastair commented on Kewl's Discussion, “CPU efficiency in Calculate node

This is probably one of the advantages of Vuo nodes being coded in C, closer to the metal. Correct me if I'm wrong, Jaymie but I always thought QC was too slow on iterating maths patches and screen draws for simple objects like GL Lines and that was why I looked forward for a C based implementation of nodes and modern GLSL implementation of rendering. Although now I don't do this kind of work much :-\

useful design's picture

Okay here's the FR: https://vuo.org/node/1723#comment-4221

Coming from this discussion around the local sub-comp FR, which is more of a virtual node kind of idea: https://vuo.org/node/1185

useful design's picture

I guess i'd go further and say the idea of this feature is simple "macro" (in QC terms) not a linked file. All the code is within the parent Vuo file. It's just for making graphs more readable and to reduce Editor screen redraw time when running the comp — especially in debug mode or with value printers showing.

Collapsed macro/Sub-Graph:

Expanded macro/Sub-Graph:

PDF of these two images which you can edit in AI if you have it installed. Sub-Graph Demo p3 copy.pdf

Video visualising and explaining the expansion and collapsing geometry with my reasoning. Sub-Graphs-ii.m4v

All this is from Sept 2011 and not particularly polished as far as node look goes, more about the expansion/collapsing. The wires have three types, values (solid line), events (dotted) and lambdas (dot/dash lines) which in 2011 was a very exciting idea/prospect that Vuo team were hypothesising about. Lambdas are common in functional programming languages and in Vuo is the concept of nodes outputting 'code' not just 'values' to be passed to other nodes which then executes the code on data (or more code!) available at other input ports.

I also worked up ideas and QC prototypes for sub-comps that would be window in window and you could pan inside the frame to see the full extent if they didn't fit all nodes in the size of box you had. I guess that means either fixed subgraph box sizes or user adjustable frame size for the macro nodes in expanded view.

useful design's picture

I definitely have asked for this (and even prototyped it in QC in pre-alpha Vuo days because the team didn't think pan within the sub-graph frame it was a good idea at the time), not sure if I made an individual FR though, will search for it.

Here's a Keynote Presentation I made on the issue, it doesn't playback the audio on more recent Keynotes (gotta love Apples work on KN don't you regressing QC comps and a bunch of other stuff), so I've exported as QT movie.

Pages